Commentary: The VMware Era is Over

Yesterday I watched the VMware Explore (née VMworld) keynote, now hosted by Chris Wolf and mostly featuring Broadcom CEO Hock Tan. New product details were light to non-existent, while the message of “private cloud first” felt 20 years out of date.

It’s almost a year since Broadcom completed on the acquisition of VMware, but in reality, it seems much longer than that. The deal was first announced in May 2022, with the shareholders of Dell profiting handsomely from the sell-off. In the last two years, many column inches have been written on the merits or otherwise of the VMware purchase. Many tears have been shed by former employees who long for the halcyon days of the early 2010s.

This year’s VMware Explore keynote perhaps gives us a true insight into the future. There is no VMware CEO figurehead making the presentations this year. Instead, we were treated to Hock Tan (not a natural orator) and head of products, Paul Turner. The end of the presentation was a rather tedious and drawn-out attempt to put the VMware platform (VCF) into the context of a fictitious company and its IT requirements.

Stagnation

In my experience, VMworld was an opportunity for VMware to present and demonstrate the next generation of its core platform, based on ESXi and vSphere. In parallel, new technologies were introduced as “Projects”. Good recent examples include Project Monterey (DPU storage and networking offload), Project Pacific (containers), and Project Ensemble (managing cloud and on-premises).

For many people (me included), VMworld was a place to meet old friends and acquaintances in the industry. So it’s understandable that so many IT professionals are lamenting the decline of VMware, even if there is no place for sentiment in the technology business.

Private Cloud

Mr Tan believes that the future is private cloud. In this piece I penned back in 2017, folks were saying that private cloud was dead. This clearly couldn’t be further from the truth, as we hear stories of workload repatriation and growth in on-premises hardware sales (here and here). The trouble is, the world is rarely black or white, but generally many shades of grey, depending on which way the wind is blowing.

Wholesale migration to the public cloud seemed like a great idea until the bills started coming in. Instead, the public cloud is now a deployment choice, rather than a strategy. Cost, sovereignty, technical choice, scale, and flexibility all contribute to the process of choosing where applications and data should best sit. For the uninitiated, this is the way things have always been. The mainframe hasn’t died off but isn’t the flagship of IBM’s business either (for example).

Does a private cloud (only) strategy make sense in 2024? If you asked me two decades ago, I would say that building a real private cloud (which wasn’t called that then) would make perfect sense. However, “cloud” should be an operating model for the delivery of resources, not a location. If you’d asked me in 2014, I would say that hybrid cloud was the way forward. I still think that today. It’s just a question of which bits go where (from an applications and data perspective).

VCF

So, was Hock Tan proposing a private cloud model? Well, not entirely. Actually, he was promoting a VMware-only model, where storage services are delivered by vSAN, containerisation is Tanzu and networking with NSX. His vision is a VMware Cloud without having to run the services himself. Naturally, other vendors (especially in the storage arena) would feel differently about this strategy, but the licensing of VMware products has changed; you get all or nothing, so vSAN (for example) must be promoted to justify the bundling price. (Side note: in the keynote, Mr Tan indicated that Broadcom had reduced 8000 SKUs to four core products. You can have any colour of Model T, as long as it’s black.)

The Architect’s View®

Last year I wrote about incremental changes at VMware Explore, prior to the Broadcom acquisition. With VCF 9 merely a twinkle in a product manager’s eye (there were no demonstration teasers in this year’s keynote), I’m drawn to think that last year’s announcements were the software flywheel effect and represented work already in progress. In contrast, this year there was nothing concrete to show, other than a strategy.

What does this mean? Potentially, VMware platform releases may be biennial in the future. Slowing development down reduces R&D costs and makes the expected savings ($8.5 billion in 3 years) more plausible. That $61 billion purchase price has to be recouped somehow.

Unfortunately, a change to product lifecycles could make VMware Explore a pretty dull experience, but I don’t expect the event to last much longer. At best it gets subsumed into a “Broadcom World”. In the short-term, there are opportunities for the likes of Nutanix, which has just posted positive financials.

We are in the post-VMware era, folks. It was good while it lasted. Anyway, we now have the AI Era in full flow to cheer us up.